Fuels to Drive Our Future (1990)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 Conversion Technologies and R&D Opportunities Various technologies can be used to convert heavy oil, tar sands, oil shale, coal, biomass, and natural gas into liquid transportation fuels. These technologies, opportunities to reduce their costs, fuel properties, environ- mental aspects, and R&D recommendations for the U.S. Department of Energy (DOE) are discussed in this chapter. More information on some of the technologies can be found in the appendixes. The technologies are addressed to the extent that they are major areas for DOE: some sections are kept brief because of their lower relevance (see Chapter 6~. The hydrogen-to-carbon (H/C) ratio of these resources must be adjusted to that of transportation fuels; pyrolytic processes remove carbon, and hydro- processing adds hydrogen. This adjustment is a major expense and con- sumer of energy in these processes. In methanol production and indirect liquefaction (Fischer-Tropsch [F-T] and methanol-to-gasoline [MTG]) proc- esses, the entire feedstock is first converted to synthesis gas (mixture of hydrogen and carbon monoxide) a major cost and energy consumption step. The next section addresses R&D for cost reduction of hydrogen and synthesis gas. The following sections address the individual conversion technologies. PRODUCTION OF HYDROGEN AND SYNTHESIS GAS Production of hydrogen and synthesis gas (syngas) is central to convert- ing fossil resources into transportation fuels. They are required to adjust the H/C ratio of these resources and to remove undesirable elements (inor- ganic, N. S. and O). Conventional transportation fuels have an H/C ratio of approximately 2 (methanol has a ratio of 4; however, from an energy view- point it can be considered approximately a mixture of CH2, the fuel, and 57

58 FUEIS TO DRIVE OUR FUTURE H2O). Coal clearly requires the largest adjustment in the H/C ratio, but even tars and higher-boiling petroleum fractions require major adjustment (Figure 4-1~. The same considerations apply to the manufacture of metha- nol or hydrocarbon fuels from biomass. Conversion of natural gas (CH4) to liquid transportation fuels, in con- trast, requires reduction of the H/C ratio. The conventional pathway, as in methanol manufacturing, is to convert the methane to syngas, the first step in hydrogen manufacturing. This syngas can then be converted directly to methanol or to hydrocarbon fuel or it can be processed further to produce hydrogen for hydroprocessing. Production by steam reforming or produc- tion by partial oxidation are the standard processes for syngas and hydrogen manufacture. These natural gas conversion processes are increasingly used in refining and chemical manufacturing. The ample world supply of low-cost natural gas indicates a continuing international engineering and R&D effort aimed at efficiency improvement and cost reduction of these processes. While there will be a continuing need for improved high-temperature materials and catalysts, there is little need for a U.S. government-supported effort on the high-temperature conversion section of these methane-based processes. When international crude oil prices are greater than about $30/barrel, the price of domestic natural gas is predicted to exceed $4/thousand cubic feet (see Appendix D, Table D-1) and production of hydrogen and syngas from coal becomes competitive. Since this is also the range of equivalent crude prices where use of coal liquefaction and shale oil are competitive, attention to reducing the cost of production of hydrogen and syngas from coal is important. Table 4-1 shows the amount of hydrogen consumed for conversion of Illinois No. 6 coal to a syncrude suitable for further refining to transporta- tion fuels. Hydrogen used for heteroatom removal is about the same or greater than that added to the liquid product, and a larger fraction of con- sumed hydrogen is found in C~-C3 gases. The total corresponds to 142 to 226 m3/barrel (5000 to 8000 ft3/barrel) of syncrude, and an additional amount would be consumed in refining to transportation fuels. An eventual com- mercial technology might achieve a 50 percent reduction in hydrogen losses to the by-products. Equipment for manufacturing this large amount of hydrogen from coal requires a major fraction of the capital investment approximately 25 per- cent and a similar fraction of total coal consumption. In coal gasification water is the primary source of hydrogen through its reaction with carbon to form carbon monoxide. This reaction is highly endothermic, and a large amount of heat is supplied by burning coal with oxygen. This can be done in a single reactor by using a mixture of oxygen and steam or by circulation of hot solids or gas plus steam through a reac-

s9 ts ~ ~ Q ,, Ct CJ) C~ _ ~ Q Ct - Ct o I o cn _ ~ cn .- o _ _ ._ os C~ o 8 ,,, ~ I =~- _ o ._ ~ ~ _ cn ._ ~ 0 o I ~ o 0 - ._ o - cn O C~ CO C) UJ o C~ I CO ~ C'') °. o o ._ ~ ._ ~ g cn C) LLJ oo ~_ :: 0 ~ · _ . . V o U. O . _ ~ LL O z o m z llJ o I 0 o o ~q o o o ca o .s ;- ca o .= o - oo o - ~: CD ~ o o

60 FUEl~ TO DRIVE OUR FUTURE TABLE 4-1 Hydrogen Sources and Uses in Direct Coal Liquefaction Pound/ Percent 100 lb Coal of Total Hydrogen Sources From coal fed 5.33 45 From H2 gas added 6.39 55 Hydrogen in Products ~ distillable liquids 8.23 70 In by-products C~-C3 hydroc~n gases 1.26 1 1 Heteroatom gases 1.24 11 Resid arid unconverted coal 0.99 8 NOTE: Data for Wilsonville Run 257H, February 1989. Illi- nois #6 coal, Buming Star Mine. Based on moisture- and ash- free coal. tor. Dilution by nitrogen is undesirable, and in the first case relatively pure oxygen must be separated from air. In the second case heat is added to the hot solids or gas in a separate vessel via combustion with air. The reaction of coal with steam produces, in addition to carbon monox- ide and hydrogen, carbon dioxide (CO2), hydrogen sulfide (H2S), methane (CH4), ammonia (NH3), and particulates. These are currently removed in a low-temperature cleanup train, and when hydrogen is the desired product the gas composition is adjusted by the reaction CO + H2O = CO2 + H2, CO2 is removed and, where necessary, CO is removed by reaction with hydrogen to form methane. The energy for water splitting, shifting, and gas cleanup and oxygen generation comes from fossil fuel combustion with additional generation of CO2. If control of CO2 emissions becomes necessary, other energy sources could be used. These include biomass, nuclear heat, and solar energy. Biomass could be used to supply heat and hydrogen by modification of technologies developed for coal. The direct use of nuclear heat will require relatively low temperature gasification and would lead to a gasification process specifically designed for integration with a nuclear heat source (see Appendix K). In the period between the oil price increases of the 1970s and the petro- leum price collapse of 1983, there was a wide variety of R&D on hydrogen

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 61 production processes with the anticipation of extensive use of these proc- esses starting in 1990. Uncertainty about future oil prices has resulted in abandonment of many of these programs. While no commercial coal lique- faction plants were built, a few other applications of gasification are provid- ing commercial experience on coal gasification in this country. Of particu- lar interest is its use in conjunction with electric power generation. The use of coal gasification for electric power generation is based on the ease of removal of pollutants from the gas and on its potential for use in more efficient combined-cycle systems where the gas is first burned at pressure in a gas turbine followed by steam generation by the hot gas tur- bine exhaust. The pioneer Cool Water demonstration program in Daggett, California, which makes use of the Texaco entrained flow gasifier, is con- sidered quite successful, and both Texaco and Shell are actively developing and marketing this type of gasifier where finely ground coal is reacted with oxygen at pressures of 150 to 600 psi in a high-velocity reactor with cocur- rent flow of coal and gas. Temperatures approach 1650°C (3000°F). Lurgi countercurrent moving bed gasifiers are in commercial use and have dem- onstrated good performance at the Great Plains coal gasification plant in Beulah, North Dakota. Assuming there is a continuing market for specific gasifiers, it is ex- pected that the continuing industrial R&D programs will leave little incen- tive for DOE to engage in research aimed at evolutionary improvements in these specific systems. While these systems have the advantages of being available and generally applicable when hydrogen or syngas is needed, they suffer from problems of durability, small scale, and inability to integrate pyrolysis tar and gas recovery with the total liquefaction system. The Texaco plants at Cool Water, for example, have a capacity of 1000 tons/day of coal. A design based on Wilsonville technology proposed use of larger-scale gasifiers (6760 tons/day [wet]~. A lOO,OOO-bbVday plant would require 12 of these gasifiers, 9 on-line and 3 standby. While mul- tiple units are needed to allow for periodic repair of ceramic liners, it ap- pears that much larger units may offer opportunity for cost savings for a scale of operation consistent with the scale needed for producing transporta- tion fuels. While these commercial systems could probably be increased in size, other systems may be more amenable to major scale-up and integration with the coal liquefaction process. The separation of oxygen from air is a major expense in current commer- cial systems, and the alternative of supplying heat from separately heated solids or gas can be competitive. The sources of heat in this case would be fuel combustion by air or, perhaps in the long term, nuclear heat. These systems, in general, would operate at lower temperatures (below ash fusion) than the oxygen-entrained flow systems. This reduces the problem of ceramic life but introduces the problem of disposing of unfused ash, which would be

62 Fuel TO DRIVE OUR FUTURE more easily leached than fused ash. Production of by-product methane increases as the temperature is reduced, and in the presence of catalysts large amounts of methane are generated in situ, as in the Exxon catalytic gasification process. Methane formation is a secondary reaction that can be limited by short gas contact time or by the presence of a CO2 acceptor such as calcium oxide. Use of fluidized bed technology appears to offer advan- tages for large-scale solids-handling processes and potential for use in multiple beds and stages. Many variations of multiple bed systems for coal gasification have been investigated (Kuo, 1984~. Figure 4-2 is a generic scheme, which provides for coproduction of hydrocarbons from pyrolysis and can supply heat for the endothermic char-steam reactions. The gasification reactor can produce large amounts of methane, if conditions are chosen to encourage catalytic gasification (as in the Exxon catalytic gasification process). Hydrogen of purity sufficient for coal hydroliquefaction could be produced, under differ- ent conditions, where circulated calcium oxide acts as a CO2 acceptor. An additional set of alternative processes involves the use of hot iron or iron oxide to produce hydrogen from steam. Iron oxide is reduced by producer gas from reaction of coal with an air-steam mixture. The reduced iron oxide (FeO) reacts in a separate vessel with steam to form hydrogen I Gas l . . J Tar Flue ' Gas , ~ 'Combustion Hot Solids aim//////; Hot Solids Coal Limestone Air Steam Gas Gasification l ~ Waste 1 Solids FIGURE 4-2 Schematic of a coal gasification process following pyrolysis and combustion to obtain higher thermal efficiencies.

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 63 and Fe3O4. A variation involving use of molten iron is being studied by the Japanese. Underground gasification of coal has been extensively studied in Russia and in the United States. Early U.S. field tests used air injection to produce a relatively low BTU product gas. Injection of oxygen and steam as gasifi- cation agents was shown in later tests to substantially improve product heat- ing values. A successful pilot scale test was carried out in a steeply dipping seam in 1981 by a joint project between DOE and Gulf Research and Devel- opment Company. This test (Rawlins II) produced good-quality synthesis gas for 65 days at a rate of 130 tons/day. The work was terminated for lack of funds (Singleton, 1982~. Estimated costs, using the test results, indicated a substantial reduction in synthesis gas cost (Schulman and Biasca, 1989) (see Appendix D, Table D-3. Because the steeply dipping coal seams qualifying for this technique are considered "unmineable," a very low coal value was used. and The most recent underground gasification field test, the Rocky Mountain I test, was conducted in 1987 to 1988 by a consortium of the DOE and several private concerns headed by the Gas Research Institute. This test demonstrated successful long-term gasification of flat seams via the use of the Controlled Retracting Injection Point (CRIP) technology developed by the Lawrence Livermore National Laboratory (Cena et al., 1988~. The Rocky Mountain I CRIP module gasified over 11,000 tons of coal in 93 days to produce a product gas with an average heating value of 287 BTU/ standard cubic foot. The gasifier operated at efficiencies in terms of energy recovery per unit of oxygen/steam injected that were comparable with sur- face gasifiers. With the potential of significant reductions in capital and mining costs compared with surface gasifiers, underground gasification is a promising technology for liquids production from coal, but several ques- tions remain, including · scaling-up in length of time and rate of gas production, · effects on aquifer quality for a large operation, · ability to predict and control performance for less ideal coal seams, · evaluation of the amount of coal meeting the requirements of this technique. A joint industry-DOE commercial demonstration designed to answer the above questions is recommended pending the environmental results from Rocky Mountain. The syngas produced from this demonstration could be used to produce ammonia or methanol.

64 FUELS TO DRIVE OUR FUTURE The DOE Program Table 4-2 summarizes the 1988 and 1989 appropriations for work rele- vant to coal gasification. Advanced Research. The programs on gasification and pyrolysis chemis- try, ash agglomeration, and gas separation are important areas. New pro- grarns with particular relevance to integrated direct coal liquefaction sys- tems should be established. Systems for Synthesis Gas Production. The systems development and modeling effort is essential to identification of the optimized gasification approach and the corresponding research thrusts. The scope should proba- bly be extended to include the total system of liquefaction, gasification, and coproducts. Systems for Coproduct Production. This large program appears to be oriented toward stand-alone pyrolysis liquid-char producing systems. No reference is made to processing pyrolysis liquids in the coal liquefaction reactor in combined pyrolysis-gasificaiion-direct liquefaction systems. As described in the 1990 proposed program, the program is aimed at directly TABLE 4-2 The 1988 and 1989 Appropriations for Work Relevant to Coal Gasification Appropriations ($1000s) 1988 1989 Relevance to Liquid Fuel Production from Coal Advanced research 2,698 2,679 High Systems for synthesis gas production 1,958 2,679 High Systems for coproducts production 5,292 9,084 Medium Total 9,948 14~442 Underground coal gasification 2,777 1,371 High Systems for power production 11,176 5,926 Low Systems for industrial fuel gas 1,369 848 Medium Great plains coal gasification (methane production) SOO 517 Low Total 15,822 8,662

CONVERSION TECHNOLOGIES ED Ray OPPORTUNITIES 65 marketable nontransportation products. The relevance to major transporta- tion fuels should be improved. Underground Coal Gasification. As discussed previously, this process has potential for reduced cost from the use of low-value coal. Production from Biomass. The 1990 DOE Solar Energy Program in- cludes $1.1 million to begin small-scale biomass gasification research. While aimed at methanol production, the same processes could be used to supply hydrogen for manufacture of transportation fuels from fossil resources. Synthesis gas or hydrogen produced from biomass has potentially no net CO2 effect on the atmosphere if this synthesis gas is used to manufacture methanol or F-T gasoline. If such production of methanol were to occur in the economy, the amount of coal-based gasoline would be less, resulting in less CO2 output. The alternative of using biomass-produced hydrogen for coal liquefaction, however, results in a larger reduction in coal-produced CO2 since more liquid fuel is produced per unit of hydrogen (synthesis gas) by the coal liquefaction route than by manufacture of methanol from syn- thesis gas. Systems for Power Production. Much of the work on systems for power production is quite specific to utility problems. Hot gas cleanup (H2S re- moval) is less important for coal liquefaction since the process is sulfur tolerant and since H2S must be removed from the spent hydrogen stream. Hot CO2 removal with concurrent shifting might be of greater interest. Systems for Industrial Fuel Gas Production. The program on systems for industrial fuel gas production aims for lower-cost oxygen production and operation of the Morgantown Energy Technology Center (METC) fluid bed gasifier. This work can potentially make contributions to gasification for hydrogen and synthesis gas production and should be managed with this . . In mend .. Conclusions and Recommendations Manufacturing of hydrogen and synthesis gas is a major economic and energy cost and source of CO2 in the production of liquid transportation fuels from natural gas and coal, shale, heavy oils, and biomass. Processes for syngas manufacturefrom natural gas are widely used and of continuing R&D interest to industry. The participation of DOE should be limited to exploratory and fundamental studies that are relevant to manufacture from both natural gas and coal. When petroleum prices increase to the point where coal and shale lique-

66 FUEL; TO DRIVE OUR FUTURE faction are competitive, it is expected that increases in the price of natural gas will cause a shift to coal gasification. Other sources of hydrogen (bio- mass, electrolysis, photolysis, thermal water splitting) are expected to be more costly. High priority should, therefore, be placed on reducing the cost of hydrogen and synthesis gas manufacture from coal, reducing CO2 genera- tion by using nonfossil sources of energy for process heat and hydrogen production, and improving energy efficiency. While general-purpose coal gasification processes are now in limited commercial use, it is believed that important opportunities exist for devel- opment of gasification processes specifically chosen for integration with direct coal liquefaction to reduce both cost and CO2 production. Such an optimized process might incorporate features such as coproduction of py- rolysis liquids and low-cost methane, larger-scale equipment, and use of air combustion, biomass combustion, or possibly nuclear heat in the long run. Demonstration and broader evaluation of underground coal gasification are also recommended. The 1989 DOE program is, in general, well chosen; however, it is recom- mended that increased emphasis be placed on identifying opportunities for reducing transportation fuel cost and CO2 output by integration of direct coal liquefaction and biomass processes in an era when natural gas prices are high. When such opportunities are identified, the program should be expanded with the goal of working toward a demonstration. HEAVY OIL CONVERSION The carbon rejection and hydrogen addition processes for heavy oils are difficult primarily because of the high concentrations of contaminants like sulfur, nitrogen, metals (mostly nickel and vanadium), and coke-forming molecules (known as "carbon residue") in the heavy oils. Metals poison catalysts and reduce upgrading efficiency, and sulfur and nitrogen must be removed in a cost-effective and environmentally acceptable manner. Many of these processes for converting heavy oil or crude oil vacuum residuum (vacuum resid) are commercial and standard in the petroleum industry. The petroleum sector is profiting from early investment and licensing of these procedures. A brief description follows (for more details, see Appendix E). Commercial Processes There are a number of commercial carbon rejection processes that up- grade heavy oil to liquids, coke, and gas, the liquids generally of a poor quality. The liquids must usually be hydrotreated before being used as reformer or fluid catalytic-cracking (FCC) feeds to make transportation fuels. These processes include the following:

CONVERSION TECHNOWGIES AND R&D OPPORTUNITIES 67 1. Delayed coking heavy oil or vacuum resid is thermally cracked in a vessel yielding liquids, gas, and high-sulfur coke. 2. Fluid coking heavy oil is thermally cracked in a reactor containing a bed of fluidized coke particles. Sulfur oxides (SO) need to be controlled. 3. Flexicoking an extension of fluid coking, in which most of the coke is gasified to low-Btu gas and the need for a coke market is eliminated. Sulfur is removed as hydrogen sulfide. 4. Resid FCC and heavy oil cracking resid is fed to a fluidized bed with a cracking catalyst, yielding gasoline-range boiling materials with car- bon residue deposited on the catalyst. Since heavy metals may poison the catalysts, upstream processing of the resid is usually required. Commercial hydrogen addition processes include catalytic or thermal hydrocracking, or the donor solvent type. They include (1) fixed-bed resid- uum or vacuum residuum desulfurization (RDS/VRDS), developed 20 years ago, is increasingly used as heavy oils become heavier. In this process atmospheric or vacuum resid oil contacts catalyst and hydrogen, removes most of the metals and sulfur, and creates an acceptable feedstock for fur- ther upgrading in an FCC. (2) Bunker flow or hycon process is similar to RDS/VRDS, except that the catalyst can continuously be added and re- moved. (3) Ebullating bed processes, known as LC-fining or H-oil, involve hydrocracking and remove metals and sulfur of any heavy oil. The distil- late products are of low quality and require further hydrotreating and up- grading. Processes with Limited Commercial Application Asphalt residue treatment (ART) is a carbon rejection process with a reactor similar to an FCC. The feed contacts a high-temperature solid and is volatized, and the coke is burned off the solid in a regenerator to produce the required heat. The liquid product yield is high but requires further upgrading. Hydrogen addition processes include many slurry hydrocracking proc- esses, a variation of thermal high-pressure hydrocracking. A dilute slurry is added to a cracking reactor to suppress coke formation and attract metal contaminants. Conversions of vacuum resid are high, but the products are high in sulfur and nitrogen, requiring further hydrogenation. Fuel Properties The product qualities resulting from the various heavy oil upgrading technologies are quite variable and are strongly dependent on feed type, process type, and processing conditions. However, producing fuels of ac- ceptable properties is possible (in all cases) with existing petroleum proc-

68 FUEL; TO DRIVE OUR Fl]TURE essing technology, although the economics vary with a given refinery's situation, the feedstock, and product prices. Environmental Considerations Air emissions can be controlled as necessary for all of the technologies (see Appendix E). All hydrogen addition processes reduce nitrogen and sulfur contaminants to ammonia and sulfur (via hydrogen sulfide), which are low-value by-products sold for fertilizer and sulfuric acid manufacture. Solid wastes consist mainly of spent catalyst. Spent resid hydroprocessing catalysts are generally quite high in nickel, vanadium, cobalt, and molybde- num and constitute decent"ores." Metals are frequently extracted and recycled. Spent FCC catalysts are suitable for landfill in most locations but in the future may require disposal as hazardous waste. The Petroleum Environmental Research Forum (PERF) is currently researching the incor- poration of spent FCC catalyst in cement, concrete, and asphalt. A few refiners already dispose of their spent FCC catalyst by that method. The wastewater treatment implications of the newer generation of heavy oil conversion processes are much more complex, and generalizations are difficult. Opportunities for Cost Reduction For well-demonstrated processes such as delayed coking or fluid coking, the potential for significant technology improvements and, consequently, for cost reductions is somewhat limited. For the other processing routes the potential for improved catalysts, equipment design, and processing condi- tions is good, and there is continuing progress in decreasing the capital investment and operating costs. In 1988 over 1.0 million bbl/day of residual fuel oil was consumed in U.S. electrical, utility, industrial, and commercial boilers. Many refineries are already equipped to convert heavy oils and residue to transportation fuels. However, most refineries use coking processes rather than the more expensive hydroprocessing technologies. Technology advances will allow more extensive use of hydroprocessing, which produces more transportation fuel than coking, especially through conversion of heavy feedstocks that are difficult to process due to high levels of metals, coke-forming molecules, and other aspects. DOE Research Program Recommendations The technical literature indicates that the Japanese government is deeply committed to funding heavy oil R&D through such organizations as Re-

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 69 search Association for Residual Oil Processing (RAROP) and the Research Association for Petroleum Alternative Development (RAPAD). It is impor- tant for DOE to fund appropriate heavy oil research if the United States is to maintain a leadership position in petroleum processing technology. Fund- ing the work in academic laboratories also serves the need to educate more U.S. scientists and engineers to ensure U.S. industrial competitiveness in the future. The committee recommends funding research on the fundamental chem- istry and kinetics of chemical reactions occurring in the existing heavy oil upgrading processes. For example, understanding the mechanisms of coke formation and how reaction pathways might be altered to reduce coking tendency is important. Such fundamental knowledge could be useful in improving many existing processes, both carbon-rejection and hydrogen- addition approaches. There is very little fundamental molecular information available about the structures of metals, sulfur, and nitrogen-binding sites and coke precur- sor species in heavy oil feeds and upgraded products. Such information would be very useful in developing new processes and improving existing ones. Such projects are appropriate for DOE funding in academic laborato- ries where expertise in sophisticated analytical instrumentation exists. For all hydrogen addition processes the cost of hydrogen is quite a sig- nificant fraction of the total upgrading cost (Fant, 1973~. The total cost of hydrogen to convert residuum to transportation fuels is in the range of $2 to $3/barrel. Although hydrogen manufacture has been researched for many years, a breakthrough in this area would greatly reduce the costs associated with many heavy oil upgrading processes. RAROP is also building a pilot plant to study inexpensive hydrogen production (Japan Chemical Week, 1986~. Further research in the environmental area may be appropriate in the areas of wastewater from newer heavy oil conversion processes. Also, improvements might be made in the extraction of metals from spent hydro- processing catalysts. The committee does not recommend DOE funding of research on process or catalyst development in the area of heavy oil conversion. There is al- ready an extensive commitment to R&D in this area in the private sector, and much duplication would likely result. TAR SANDS RECOVERY AND PROCESSING Tar sands are defined as "any unconsolidated rock containing a crude oil which is too viscous at natural reservoir temperatures to be commercially producible by primary recovery techniques"; American Petroleum Institute gravities are generally less than 10° (IOCC, 1982; see also Appendix C).

70 FUELS TO DRIVE OUR FUTURE While U.S. tar sands deposits are not as extensive as Canada or Venezuela, they are significant (see Table Cub. U.S. tar sands are hydrocarbon wetted rather than water wetted like Canadian tar sands. The percent by weight of bitumen of the U.S. resource varies widely, from about 1.5 in Alabama, to 4 to 10 in Utah, to 30 in California—a value that essentially represents heavy oil (Tables 4-3 and 4-4~. Porosity, a characteristic of the host rock, ranges from 15 to 40 percent of total rock weight, and oil saturation usually repre- sents about 50 percent of the porosity. Overburden, which varies by site, is an important determinant of the total cost of tar sands recovery, as is the extent of layering of a given deposit, and bitumen properties also vary widely among different deposits. Recovered hydrocarbons would generally need some onsite viscosity reduction to make them pumpable through refin- ing systems. Like heavy oils, high-sulfur bitumen results in high-sulfur coke in conversion processes producing coke requiring SOx emissions con- trol. Nitrogen content, generally higher for tar sands bitumen than petro- leum, can poison refinery catalysts and affect end-product fuel quality. Conversion of tar sands to products requires mining, recovery, and upgrad- ing. TABLE 4-3 Tar Sands Reservoir Characteristics Grade Porosity Oil Overburden Thickness (% oil) (%) Saturation (%) (ft) (ft) Utah 4-10 15 23-72 0-500 10-60 Texas 30 35-55 1500 15-300 Kentucky 6-10 15 20-70 0-200 10-50 California (up to 30) 30-40 50-75 0-3200 50-400 Alabama 1.5 6-24 4-56 0-1000 10-300 TABLE 4-4 Bitumen Properties Viscosity API Pour Point (million Sulfur Nitrogen Gravity (OF) centipoise) (wt%) (wt%) Utah 5-14 95-150 1 0.4-3.8 0.6-1.3 Texas -2-10 180 20 10.0 0.4 Kentucky 10-12 55 1 1.5 0.4 California 8-17 1 3-7 1.2

CONVERSION TECHNOLOGIES ACID Rid OPPORTUNITIES 71 Mining The costs of surface mining recovery of tar sands ore depend on the ore body configuration, thickness, and type and the overburden of nonbearing rock. Overburden thicknesses greater than around 300 to 500 ft result in large initial costs for tar sand commercial plants and significantly reduce economic viability. Variation of ore richness within a deposit affects the mining strategy and production costs. Mining techniques include drag lines and bucket wheel excavators, which are used in Athabasca tar sands in Alberta, Canada, and are quite effective for bulb surface mining of unconsolidated tar sands. Mining of consoli- dated tar sands, such as those of tar-saturated sandstones, requires drilling, blasting, and removal with power shovels and trucks. Commercial convey- ors are available to move either type of tar sand to the processing plant. Newer mining machines, similar to road resurfacing machines for asphalt roads, are mounted on Beads and use a rotating drum fitted with tungsten carbide teeth to rip out thin layers of tar sand ore or overburden. Mining machines are particularly suited to multiple seams of consolidated tar sands. Recovery Two types of recovery technologies, extraction or retorting, are generally considered for tar sands. In extraction a solvent such as naphtha dissolves bitumen from the host rock at low temperatures. For retorting tar sand is heated to pyrolysis temperature in a retort vessel. For U.S. tar sands extrac- tion is more cost effective than retorting and is the preferred technology. A higher oil-yield retorting process might make retorting more economical for deposits with lower ore richness. Extraction is more cost effective than retorting for tar sands below about 15 percent by weight richness. This richness includes most of the U.S. resources. Extraction Depending on the type of tar sand material, either a water-based or hy- drocarbon solvent is needed to recover tar sand bitumen from host rock. The Clark hot-water process, using a caustic water solution to emulsify oil from the tar sand particles, is effective for Canadian tar sands because the sand particles are wetted with water and surrounded by bitumen. The proc- ess recovered initially about 85 percent of the bitumen. Recent process im- provements have increased recovery to between 88 and 92 percent (Coal and Synfuels Technology, 1989a). The unrecovered bitumen is discharged with the tailings, causing a serious environmental concern. U.S. tar sands, in contrast, are hydrocarbon wetted, and effective bitu-

72 FUEI~; TO DRIVE OUR FUTURE men recovery requires a hydrocarbon solvent used in processes specifically tailored for a given process. Figure 4-3 shows a generalized solvent extrac- tion process. Tar sands and solvent come together in an extraction contac- tor, such as a mix tank, rotating drum, or other mixing device. In this example water is also added to wet the mineral particles after the solvent has had time to wash bitumen from the sand. Separation of the solvent- bitumen phase from the water-sand phase is easier than in the hot water process, since emulsions can generally be avoided with the solvent process. Tailings can be separated in a raked bottom separator by gravity. Using water in this process to displace solvent and bitumen from sand saves en- ergy in recovering solvent from the solid tailings. The bitumen solvent phase is separated, and fine solids are removed in a centrifuge, lamella, or by other means. Solvent recovery is accomplished in evaporators or by distillation to produce a relatively solids-free bitumen. Recycled solvent is recovered for reuse. Critical features of hydrocarbon solvent extraction i] ,nclude essentially complete recovery of solvent from spent tar sands and good removal of mineral fines from bitumen. Residual solvent concentrations of only 100 ppm represent substantial losses of solvent, causing environmental prob- lems and reduced profitability. The extent of fine solids removal affects the adaptability of existing refinery processes for upgrading bitumen to lighter products. Solvent Tar Sands ~- Recycle Solvent Water r Recycle Solvent l 11 1 Fines Solvent Extraction ~ Removal ~ Recovery Tailings Fines Bitumen FIGURE 4-3 Hydrocarbon solvent extraction process.

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES Retorting 73 Many possible retorts might work for tar sands, but the Taciuk, Allis- Chalmers, Lurgi, and fluid-bed retorts have been the subject of the most recent R&D. The Taciuk retort is under development in Canada by the Alberta Oil Sands Tar Research Authority (AOSTRA) for Canadian tar sands but can be used on U.S. hydrocarbon-wetted tar sands. The Taciuk retort is a rotating drum device, 9 ft in diameter, with efficient, countercur- rent heat transfer, which effectively retorts and cracks tar sand to light oils and gases; the feed material is predried by contact with exiting combusted spent sand. A 9-ft-diameter pilot retort has been tested to date. A similar rotating drum retort was under development some years ago by Allis-Chalm- ers near Milwaukee. The Lurgi-Ruhrgas retorting process has been commercially used for coal pyrolysis. When using the process with tar sands, the feed is mixed with hot combusted spent sand in a screw conveyor and added to the retort- ing reactor, which operates at 480° to 510°C (900° to 950°F) (Figure 4-4~. The bitumen cracks to gas and light oil, which pass overhead as a vapor, and to coke, which deposits on the spent sand. A cyclone separates sand fines from the product vapors, and the vapors are then condensed to an oil and water stream. The retorted sand flows downward to the bottom of a riser combustor pipe, where air is injected to burn the carbon and unreacted bitumen. The combustor reaches temperatures of 590° to 650°C (1100° to 1200°F). Sulfur and nitrogen emissions in flue gas must be properly con- trolled. Spent sand can be rejected from the product cyclone or retort as shown, but the preferred configuration is to withdraw and cool spent com- busted sand from the cyclone. Fluid bed retorting processes also provide residence time for retorting and burn coke from the spent sand to achieve environmentally acceptable tailings. Upgrading Upgrading bitumen to higher H/C ratio products is similar to upgrading heavy oil, either through carbon rejection or hydrogen addition processes. The choice of process depends on oil prices, bitumen reactivity, sulfur, nitrogen content, and the combination of upgraded products desired. In general, crude products from bitumen upgrading or retorting conver- sion processes are further upgraded to produce liquid feedstocks for exist- ing refinery units and by-product fuel gas. The naphtha product is hydro- treated to remove nitrogen prior to further refining, to produce high-octane motor gasoline blending stocks. The distillates also need to be hydrotreated to meet diesel, heater oil, or furnace oil specifications. Gas oil is hydro- treated to remove nitrogen so it can be fed to a catalytic cracker ("cat cracker"), which further cracks the gas oil to additional naphtha, distillate,

74 FUELS TO DRIVE OUR FUTURE Sand ~ l Sands clone iL v Moving Bed o Retort ~ / )? ~ Al Moving Bed Retort Flue Gas }4 }I Gas ~ ~ ,_. _ A ~ I _ vvalur Cyclone / / FIGURE 4-4 Lurgi-Ruhr gas process. Spent Sand and gas. Since sulfur is more easily removed than nitrogen in hydrotreating, the sulfur content of upgraded products from Utah or other low-sulfur tar sands will be very low. Environmental Considerations Tar sands processing converts part of the bitumen to coke and fuel gas, which are burned in processing plants and refineries to generate steam and power. Sulfur and nitrogen oxide emissions controls are not expected to be significantly different than those for present refineries. Sulfur oxide con- trol should, in fact, be somewhat easier for very low-sulfur Utah tar sands. Likewise, environmental controls for water quality related to mining

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 75 operations do not present any new or special requirements over current practices. On the other hand, new techniques are required to ensure that the spent tailings from retort or extraction operations can be disposed of safely in a landfill. Complete combustion of hydrocarbons on retorted sand will be required. For extraction processes the complete removal of leachable hydrocarbons will be required. Where a solvent is used, complete recovery of the solvent will be critical. The consumption of fuels made from low-sulfur Utah tar sands should produce lower sulfur oxide emissions, since average refinery products have higher sulfur levels. Nitrogen emission levels are expected to be about the same as for current refinery products. Economics The overall economics of a commercial tar sands plant depends on the following, in descending order of importance: (1) resource richness, since richer ores require less ore to be mined and are easier to process through primary recovery; (2) bitumen quality, primarily its nitrogen and sulfur content as well as its type and resulting yield structure from upgrading; (3) reduction of energy process requirements, to avoid purchasing large amounts of electrical power, on the one hand, and expensive capital outlay for power generation equipment, on the other; (4) such resource characteristics as overburden thickness and the ability to mine selectively high-grade portions of deposits; and (5) resource location as it affects markets, transportation, and labor productivity. Using extraction technology and bitumen upgrading through an asphalt residue treatment process, the costs for upgraded tar sand products are about $26/barrel to $35/barrel. The costs for pyrolysis processes are generally higher. A number of cost reductions for U.S. tar sands processing could improve its overall economics. These opportunities include improved bitumen up- grading; higher bitumen recovery, which would require less mining; im- provements in solids removal, which could amount to 15 percent of the total cost of products depending on mineral characteristics; and improved tech- niques that permit more selective mining of the richest ore, which would reduce both processing and mining costs. The committee judged that these improvements and the use of a rich resource such as Utah tar sands could reduce costs to about $20/barrel of crude equivalent (based on the 10 per- cent discount rate cited in Chapter 3~. Recommended Areas of Research for DOE Based on the above considerations, a broad R&D program for exploiting domestic tar sands should include the following areas:

76 FUELS TO DRIVE OUR FUTURE · characteristics of nitrogen compounds, to improve the upgrading of low-API-gravity, high-nitrogen bitumen; · solid-liquid interactions, to improve removal of fine solids; · dissolution mass transfer effects to optimize the extraction and separa- tion operations; improved extraction techniques, to improve bitumen recovery; upgrading of feeds containing solids (need pilot plant demonstrations) due to solids handling requirements for disposal; environmental studies, especially for solid waste disposal; and improved techniques for efficient and selective mining of the richest ores. The current DOE program for surface extraction of tar sands is focused on the Western Research Institute (WRI) bitumen recycle process and opti- mization of a few alternative recovery processes. Over the next 5 years these and other available processes along with mining techniques applicable to U.S. tar sands should be evaluated both technically and economically. Engineering firms, petroleum operating companies, and qualified consult- ants should play a key role in these evaluations, which should be used to determine if any processes are suitable for further development in a field pilot facility having a capacity of 50 to 100 bbVday. Building on Canadian experience, this size should be suitable for scale-up to a commercial plant. A field pilot operation is justified only if the technology is judged to be sound, all enviromental requirements are projected to be met, and costs are sufficiently low (probably below $25/barrel) to attract industry participa- tion. If a technology worthy of development cannot be identified, serious consideration should be given to terminating the government program. Continuation of the program would be justified only if the prospects for improvement were judged to be outstanding. OIL SHALE The Green River formation in Colorado, Utah, and Wyoming is the larg- est and richest oil shale deposit known. The most valuable part of the resource is in the Piceance Creek basin in western Colorado. The northern and central parts of the basin (600 square miles) contain about 720 billion bbl of shale oil (Lewis, 1980) in thicknesses ranging from 400 to over 2000 ft with a relatively thin overburden. Almost all this thick, rich shale is owned by the government, the main exception being the two federal leases (tracts C-a and C-b) held by Amoco and Occidental Petroleum, respec- tively. Much of the resource is suitable for very large scale, low-cost open pit mining. This possibility has received little attention, perhaps because most of the land is government owned and is not available for lease or sale. Tract C-a is suitable for such open pit mining.

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 77 A smaller area in the southern part of the basin is privately owned, mostly by oil companies. The oil shale of economic interest in this area consists of a thin, high-grade layer of shale, the Mahogany zone, generally 30 to 150 ft thick. This layer of oil shale, averaging 30 gaVton' contains approximately 50 billion bbl of shale oil (Lewis, 1980~. This layer is suitable for room and pillar mining but generally not open pit mining be- cause of the high ratio of overburden to oil shale. In the central part of the basin the Mahogany zone is thicker and richer. The other zones, especially those deeper in the section, also contain economically important quantities of oil. The Green River formation also underlies large areas of Utah and Wyo- ming. In these areas the shale that averages more than 20 gal/ton is less than 400 It thick. High-grade layers in these regions are for the most part too thin or too deeply buried to be of economic interest for the foreseeable future, although in a small area of eastern Utah shale occurs close to the surface. Although the shale in this area is generally thinner and of lower grade than that in the southern part of the Piceance basin, it is otherwise similar and suitable for room and pillar mining. Deposits of oil shale other than the Green River formation occur in the continental United States and Alaska. The most widespread of these are the oil shales of Devonian to Mississippian age, generally referred to as eastern shales. These eastern shales are thinner and of lower grade than the Green River formation as measured by Fischer assay, a standardized test that meas- ures the amount of liquid oil that can be obtained in ordinary pyrolysis processes. Kerogen in eastern shales has a lower H/C ratio and is more like coal than the Green River shales. For thicknesses of 10 to 30 ft the average grade is usually between 5 and 10 gal/ton by Fischer assay. Increased conversion to liquids by adding hydrogen is possible but requires a more complicated and expensive process (Hu and Rex, 1988~. The following sections will concentrate on the shale of the Green River formation, which has by far the most economic potential. State of Technology Development Shale Properties and Process The Green River oil shale is an impure marlstone (a silicate/carbonate rock) consisting mostly of inorganic mineral matter. The solid organic constituent is kerogen, in a typical shale making up about 15 percent of the rock by mass and 30 percent by volume. Due to the nature of the oil shale, any practical physical process to completely separate the kerogen from the mineral matter is unlikely. A less than complete separation might be useful if the beneficiation process is sufficiently inexpensive. Because the raw

78 FUEl~; TO DRIVE OUR FUTURE shale is essentially impermeable to fluids and the kerogen is not soluble, ordinary (subcritical) liquid extraction processes are likely to be unsatisfac- tory. Pyrolysis or retorting is therefore the process of choice. In surface retorting shale oil is broken into pieces small enough to permit good heat transfer. (In any shale mining and crushing operation this re- quirement is easily met, although the size range and content of fine par- ticles present some handling difficulties.) The shale pieces are then heated to about 930°F (500°C) to pyrolyze the kerogen, producing oil, gas, and solid char. Heat for the retorting process is provided most efficiently by burning the solid char formed on the shale mineral during retorting. Carbonate minerals, dolomite and calcite, are major components of these shales, and dolomite especially may be decomposed at high temperatures. Decomposition of carbonates during processing wastes energy and produces CO2, which may be desirable if there is a market for CO2 and undesirable if it proves necessary to limit the production of greenhouse gases. Some reaction of carbonates with SO2 is desirable to eliminate the emission of sulfur gases. Carbonate decomposition may be avoided at practical retort- ing temperatures, but some decomposition usually occurs at the higher tem- peratures used for char combustion. The extent of carbonate decomposition depends on the time and temperature required for combustion. Shale oil produced by current processes is somewhat unstable, has a high viscosity and a moderate dust content, and usually cannot be put directly into a pipeline. Upgrading has been required to reduce the viscosity and stabilize the oil for transport to refineries. If a hydrotreatment process is used for this upgrading, the sulfur and nitrogen contents are also reduced and a premium refinery feedstock is obtained that contains almost no heavy residual fractions. The challenge is to devise a simple and efficient oil shale process for mining, retorting, upgrading, and disposal at low cost. Reaching this goal will require R&D and translation to commercial practice. Mining and Disposal To a great extent mining technology is transferrable to oil shale. The scale of the operation is large, however, and there are opportunities for improvements. Room and pillar mining on a large scale have been demon- strated in the Anvil Points, Colony, and Union mines in the Piceance basin. Much of the larger resource of rich shale in the central Piceance basin is ideal for open pit mining, with very low stripping ratios (overburden/ore). This mining would be on a large scale because of the thickness and extent of the resource. Open pits would be 2000 to 3000 ft deep and thousands of feet across. Disposal of waste is included with mining because of similar materials handling and the necessity for an integrated operation.

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 79 The ideal process would permit consolidation of waste shale, other solid wastes from the plant, and wastewater. While studies indicate that this approach is feasible, disposal of these materials in an open pit has not been demonstrated. Both spent shale and mined but unprocessed rock (overbur- den and lean shale) must be deposited in the pit and compacted in a stable configuration. Additional space outside the pit would also be needed, to the extent that the waste solids will not pack as densely as unbroken rock and allow working space during early development of the mine. For the still large remainder of the resource in the central Piceance basin, where the overburden stripping ratio is too high for open pit mining, an- other mining method must be found. Room and pillar mining can be used, but as the depth and thickness of the shale increase, the fraction of the resource that is recovered will be less than 50 percent. No underground mining method is now available for oil shale that can recover a larger fraction of the resource. Block caving is a low-cost mining method that has proved successful for massive copper ores. Large-scale block caving, if applicable to oil shale, could result in substantial cost reduction and im- proved resource recovery. In these deep zones modified in situ processing (like the Occidental Petroleum Process) is a possible way of recovering this resource and may compete favorably with room and pillar mining. Retorting Different retorting processes have evolved from different methods for heating solid particles of oil shale. Either hot gas or hot solid material may be used to supply heat to the shale. (Hot liquids are not practical for tem- peratures near 500°C.) If hot gas is chosen as the heat transfer medium, the crushed shale is distributed in a bed through which hot gas is pumped. The various types of hot gas retorts have certain characteristics in common. The mass of gas required to process a unit mass of shale ranges from 0.6 to 0.8, owing to the relative heat capacities of the shale and gas. As a result very large volumes off gas are required, which is undesirable because of the expense of com- pressing gas and the processing time required for gas-solid contact. Fine particles in these packed bed retorts restrict the flow of gas or increase the pressure drop across the bed. Therefore, shale particles with diameters less than about ifs in. must be discarded or processed in some other way. Large particles, on the other hand, require a relatively long time to heat, adding costs because of the increased residence time of shale in expensive equipment and the loss of oil through coking, which occurs if the heating is very slow. The practical size limit for shale particles in hot gas retorts is probably a diameter of between 2 and 3 in. The second option for heating the oil shale is to use a hot solid material.

80 FUEl~i TO DRIVE OUR FUTURE The raw shale is crushed small enough for rapid heat transfer. Either shale itself (after retorting) or another solid may be used for the heat-carrying medium. The solid heat carrier is heated by burning the residual carbon on the spent shale. Appendix F contains a brief description of several hot gas and hot solid retorting processes. Upgrading The properties of shale oil vary as a function of the retorting process. Fine mineral matter carried over from the retorting process and the high viscosity and instability of shale oil produced by present retorting processes have necessitated upgrading of the shale oil before transport to a refinery. After fines removal the shale oil is hydrotreated to reduce nitrogen, sulfur, and arsenic content and improve stability; the cetane index of the diesel and heater oil portion is also improved. The hydrotreating step is generally accomplished in fixed catalyst bed processes under high hydrogen pres- sures, and hydrotreating conditions are slightly more severe than for compa- rable boiling range petroleum stocks, because of the higher nitrogen content of shale oil. Shale retorting processes produce an oil with almost no heavy residual fraction. With upgrading, shale oil is a light boiling premium product more valuable than most crude oils. Considerable cost savings would be possible if the oil could be directly transported to a refinery. Hydrotreating costs would be reduced and dupli- cation of operations would be eliminated by using existing facilities. Ex- panding or modifying refinery components where necessary would be less expensive than building new facilities in the field, and the operating costs at existing refineries are generally lower than the costs at remote locations. New retorting processes may permit this option. Advanced Retorting Technologies New aboveground shale technologies are needed to reduce residence time, increase oil yield, improve oil quality, and thereby cut costs. Efforts to develop modified in situ (IRIS) retorting aim to cut costs by reducing the mining expense as compared to conventional retorting. If MIS retorting is successfully developed, the shale mined will be processed in aboveground retorts. In general, hot gas retorts are current technology, and hot solid retorts are advanced retorting technologies (Table 4-5~. In addition, hot gas retorts using internal combustion produce a low fuel value gas and have high mineral decomposition compared with high fuel value gas and low mineral decomposition with either hot solid systems or hot gas retorts using external combustion. For hot gas retorts the shale is crushed to 1~/2 to 3 in., with shoe less

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES TABLE 4-5 Comparison of Hot Gas Versus Hot Solid Retorting Systems Hot Gas Retorts Hot Solid Retorts Commercial operations High throughput of gas Slow shale heat-up Large shale required Fines discarded Sulfur released as hydrogen sulfide Oil yield below Fischer assay No commercial operations Low throughput of gas Rapid shale heat-up Smaller shale required Fines processed Sulfur retained as sulfates Oil yield at Fischer assay 81 than i/e in. rejected. Hot-solid retorts require a top size of i/. in., with all shale including fines used in some process schemes. Hot Recycled Solid (HRS) Processes. Virtually all proposed advanced oil shale retorting systems use hot recycled shale as the heat carrier, provid- ing rapid mixing with raw shale, rapid heating, and a subsequent soak time of 1 to 2 min for pyrolysis to occur. This process greatly increases through- put and reduces costs, compared to hot gas processes. An HRS process uses the following major components: a raw and recycled solids mixer, a soak tank pyrolyzes, a pneumatic transport, and a combustor. Many mixer types have been proposed, including fluid beds, gravity fall units, and vari- ous mechanical mixers. A requirement for both mechanical and gravity fall units is very rapid mixing (in less than 10 to 20 s) in a compact space (high product space velocity), to avoid excessive oil loss by cracking. In fluid beds longer mixing times are acceptable because oil vapor is quickly swept from the bed. However, more difficulty in condensation of the oil vapor and loss of fines by elutriation are disadvantages of this method. The soak tank pyrolyzer can be either a moving packed bed or a fluid bed. In some designs the mixer and pyrolyzer are combined into a single fluid bed unit. Disadvantages of a fluid bed pyrolyzer are the expense of gas compression and cooling and the problem of nonuniform and nonop- timum solid residence time. Experiments have shown a moving packed bed pyrolyzer to produce a lighter, less viscous oil that is relatively stable after cooling (Cena and Mallon, 1986~. This method may allow direct transport to a refinery without field upgrading, a major cost reduction. In addition, moving packed beds have the advantage of processing essentially all the shale because fines remain in the bed and are retorted. An air pneumatic lift pipe transports the shale upward to the combustor. A lift pipe high enough to complete combustion is difficult to operate with

82 FUELS TO DRIVE OUR FUTURE expected variations in shale grade and reactivity so it may be desirable to combine a fluid bed combustor with a lift pipe. Another system uses a delayed-fall combustor downstream of the lift, which retards the downward fall of solids while air is blown upward or crosswise, providing in a short height the required residence time (10 s) for combustion. Information from process development or pilot-scale units is not avail- able. Chevron briefly tested a fluid bed IS process at a scale of 150 tons/ day at its Salt Lake City refinery in the mid 1980s. Exxon also fielded a pilot test of its own design at its Baytown refinery during this time. Results of both tests were rumored to have been successful but are not publicly available. The largest tests of the process in the public domain were con- ducted from 1984 to 1988 by DOE at the Lawrence Livermore National Laboratory (LLNL) using a 1-ton/day solid recycle retort operated from 1984 to 1988. These experiments allowed study of chemical reactions im- portant in the process and tested both a fluid bed and gravity bed pyrolyzer and a lift pipe and delayed fall combustor system. Substantial differences in oil properties were observed between these two pyrolyzers, with major improvements in oil properties observed in the gravity bed pyrolyzes. Re- actions occurring between the gas and solids (some catalytic) convert gas to liquid, crack the liquid, and reduce the viscosity of the shale oil (without loss in liquid yield). These reactions are not yet well understood, but the possibility exists to produce shale oil that can be transported to a refinery without upgrading. The LLNL retort is being modified to process 4 ton/ day, which will allow tests using the full particle size range of commercial plants and permit study of some important solids handling questions. Environmental Considerations Advanced retorting techniques, using hot recycled solids, offer distinct advantages for controlling gaseous emissions. Pyrite forms hydrogen sul- fide in the pyrolyzer in both hot gas and hot solid retorts. However, Fe2O3 in recycled shale scavenges hydrogen sulfide (H2S) in hot solid systems, reducing H2S concentrations to under 1000 ppm. In the combustor FeS burns to form SO2. Here carbonate minerals scavenge SO2 at combustor temperatures below 700°C, and at these low temperatures nonfuel NOX emis- sions may be avoided. NOX emissions from the fuel are low. Further reduc- tion may be required and may be possible with additional study of the nitrogen chemistry. More complete combustion may be necessary to reduce CO emissions. In any case the quantity of gas requiring cleanup is gener- ally much less in hot solid than in hot gas retorts. Use of any fossil fuel contributes to release of CO2 (see Chapter 5~. The amount of CO2 released per megajoule of useful power is process dependent and varies widely. Western shales contain an average of 40 wt% carbonate

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 83 minerals, which will decompose if held too long at elevated temperatures. Hot solid processes can be designed to minimize the time that carbonate minerals see elevated temperatures. Mathematical model calculations and small pilot experience indicate that carbonate decomposition can be held below 10 percent in a typical HRS process, thus reducing CO2 formation. Waste shale as it comes from the retort must be cooled, transported, and disposed of in the mine or elsewhere. In concept the waste will be cooled, moistened, and compacted into a strong impermeable mass that will stay in place and not contaminate groundwater. Waste shale with part of the car- bonate decomposed is a natural cement and will set up after disposal, form- ing a low-permeability material with adequate strength. Experience in both wastewater treatment and shale disposal awaits devel- opment of process schemes and demonstration tests. Unocal has demon- strated successful methods and experience in these areas for hot gas proc- esses. Similar data are needed for hot solid processes on a comparable scale. Less direct environmental impacts include the growth and development of the area, with more people, more activity, more roads, and more pollu- tion. The construction of mines and processing facilities will not disrupt other human works, because the area is largely uninhabited; but it will affect natural scenery and wildlife. Development of water resources and use of water are always important issues in the western states. Shale oil production of 1 to 2 million bbl/day is possible without importing water into the basin (Brown and Stewart, 1978; Sparks, 1974~. Larger-scale production would probably require water to be imported but at a cost estimated to be a small part of the total cost of shale oil production. Requirements for prevention of significant air quality deterioration in some areas will also be important. The present prevention of significant deterioration requirements, both federal and state, particularly in Colorado, could limit the production of oil from shale. Potential for Cost Reductions in Oil Shale Processes Numerous studies have been conducted to assess the profitability of pro- ducing oil from shale. Large variations in the results depend on assump- tions about end use, scale, resource type, equipment, contingencies, and risk factors. The endogenous price calculation in Chapter 3 for oil shale indicates a price of about $43/barrel of crude oil equivalent. Potential cost reductions in the major categories involved in oil shale production and conversion are given in Table 4-6, not including those accruing from making it possible to transport shale oil to a refinery by pipeline without upgrading. Economic studies of both surface retorting and of combining MIS with HRS above

84 Fuels TO DRIVE OUR FUTURE TABLE 4-6 Potential Cost Reductions for Oil Shale Conversion Improvements in Current Technology Total Capital Total Product Cost Reduction Cost Reduction Range (%) Range (~o) Developing Advanced Technology Total Capital Total Product Cost Reduction Cost Reduction Range (%) Range (%) Mining 2-6 3-5 2-11 3-9 Retorting 3-6 3-4 8-20 5-15 Upgrading 2-4 1-3 8-13 6-10 Environment 1-2 1 2-3 2-3 Total Cost- Savings Potential 8-18 8-13 2047 16-37 surface retorting systems indicate costs in the high $20s to high $30x/barrel of oil produced (U.S. DOE, 1989c; Piper and Ivo, 1986~. These results, coupled with the possible cost reductions in Table 4-6, indicate that a cost target of $30/barrel or less is possible with development of advanced tech- nology. Recommendations for DOE Research Program for Oil Shale Development The annual DOE program in oil shale is about $10.53 million for fiscal year 1989 with both eastern and western oil shale included. Of the total, $7.46 million is for technology base studies with about $2.6 million of this for eastern oil shale. Environmental mitigation studies constitute about $3.07 million of the budget. Development of oil shale requires substantial lead time and steady progress toward demonstration of promising technolo- gies. The private sector has greatly reduced efforts to develop oil shale technology in favor of developing and producing petroleum and protecting short-term profitability. The major oil companies in particular perceive better near-term opportunities in foreign exploration and production. Gov- ernment ownership of the thickest and richest oil shale resource is also a factor in discouraging private investment. Few if any companies have suf- ficient resources to justify more than one or two plants, which probably could not realize the economic benefits of improved technology. At the same time it is not feasible for government to sell or lease substantial parts of its holdings until an industry is started and a value established. Govern-

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 85 ment involvement is therefore essential to advance oil shale technologies to the next stage of development. Technology development in this field will benefit most from the appro- priate mix of fundamental and applied research, process development and scale-up on a step-by-step basis, and industrial experience in operating plants that use full-scale components of all process steps. Much of the knowledge gained in the past has been lost as efforts were halted, yet much expertise remains to provide a base for a renewed effort. Research Areas Materials Handling and Solids Flow. The major obstacle to efficient scale-up in the oil shale industry is the handling of bulk solids. Throughout industry, processes that handle solids are prone to long start-up times and operating problems. Fundamental studies of granular flow and materials handling are needed to reduce the costs of oil shale processing. Reactions and Kinetics of Kerogen Retorting. More work is needed to understand the fundamental chemistry of oil shale retorting, especially spe- cific reactions and kinetics, to solve problems and reduce costs. Additional work is needed in pyrolysis and combustion chemistry to better design and optimize the retorting process and equipment. Vapor and mineral interactions in the retort affect oil yield and such oil properties as viscosity and stability. Field upgrading might be eliminated if the coking and cracking reactions between oil vapor and solid surfaces are better understood and shale oil with low viscosity and stable enough for pipeline transport can be produced. Nitrogen and sulfur chemistry are important in reducing NOx and SOx emissions and also in cooling of waste shale with water and eliminating hydrogen sulfide release. Decomposition of some carbonate minerals and cement-like reactions occurring during and after disposal are important to efficient spent shale disposal and prevention of leaching of waste and con- tamination of groundwater and surface streams. Process Application Chemistry. Process chemistry of waste cooling, waste solid consolidation, and leaching is important to reduce process costs and prevent environmental damage. Studies are needed of reactions be- tween water and minerals during cooling, disposal of solids and wastewater, and exposure of waste solids to groundwater. Removal of fine mineral matter from product oil is the first step in upgrading. Upgrading of shale oil is usually a process applied to oil after retorting. It may also occur as a result of the primary conversion process in the retort. Research on advanced upgrading processes should be done.

86 FUELS TO DRIVE OUR FUTURE Improved methods for removal of solids from vapor and liquid oil are also needed. Advanced processes for downstream upgrading include better hy- drotreating catalysts, and catalysts, absorbents, and solvent extraction or other means for nitrogen removal. Methods are also needed to stabilize shale oil and reduce its viscosity to allow direct shipment to refineries. Modified In Situ Processes. Research applicable to MIS processes in- cludes gas cleanup, CO2 removal from the gas stream (if reducing CO2 emissions becomes important), and understanding and preventing the con- tamination of groundwater from mining or leaching of spent shale in under- ground retorts. Mining Research. Room and pillar mining and open pit mining are adequate for a large part of the oil shale resource during early development. In the longer term a method of underground mining, perhaps some kind of block caving, should be developed to recover a large fraction of the oil shale in the north-central part of the Piceance basin that is too deep for open pit mining. Research on groundwater contamination by mining also is important. While laboratory experiments and calculations (including mathematical models) are essential, in many cases the important process phenomena can be detected or studied only in the field on a larger scale. Research must therefore be conducted in the field in pilot plants as well. Advanced retort- ing technologies (hot recycle solid processes) are ready for small-scale field pilot plants. Solids handling, gas isolation methods, continuous operation, proof of concept, materials and durability, scale-up design, and cooling and disposal of waste shale can be studied at this scale. Timetable for Development of Oil Shale Technology Oil shale will be a potential source of large amounts of liquid fuel for a very long time. Improvements in technology and industrial experience can make the resource competitive. A long lead time is required to develop technology and to obtain substantial production capacity. Development of advanced technology for a demonstration plant could be accomplished in a decade. From the decision to proceed, 2 years would be needed to design and construct a field pilot plant or process development unit (perhaps 100 bbVday) for surface retorting. Two more years would be required to oper- ate the plant and plan a full-size pilot plant (1000 bbl/day). An additional 5 years would be required for construction and successful operation of the full-size pilot plant and design of a demonstration plant. With this time- table the United States could have the option by the year 2000 to proceed with construction of a demonstration commercial module (5000 to 10,000

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 87 bbVday) that could take 3 to 5 years from construction to operation. This will demonstrate the technology and allow some cost reduction through op- erating experience. Further cost reduction can then be expected with con- struction and operation of a first commercial plant, which would take an- other 3 to 5 years. Thus, the United States could be in a position to begin to produce shale oil with low-cost advanced technology by 2010. Demonstra- tion of combined in situ (MIS) and surface retorting technology takes the same amount of time, with the critical path being surface retorting. Based on the scenarios presented in Chapter 1, the cost of crude oil could exceed $30/barrel before an advanced oil shale technology is demonstrated. Because of the long lead time required, government ownership of the land, and industry attraction to short-term opportunities for petroleum ex- ploration and development overseas, industry is not willing to aggressively develop this technology at present. It seems appropriate for the government to take the lead in its early development. However, it is important for industry participation to grow at each scale-up step. Successful demonstra- tion will require the full participation of industry on a cost-shared basis. The decision to proceed with each step will also depend on the success of the technology in cutting costs and a reevaluation of future petroleum prices. The option to proceed with a demonstration plant by the year 2000 requires an accelerated R&D program in comparison to the current one. The com- mittee also judges that, because of the nature of the resource, development of eastern shale should not be conducted at this time. SYNGAS-BASED FUELS Syngas can be converted to liquid fuels in two ways: (1) conversion to methanol, which can be followed by conversion to gasoline and distillate, and (2) via the Fisher-Tropsch (F-T) process to hydrocarbons. Methanol from Syngas Methanol is made by the catalytic conversion of syngas at about 250°C and 60 to 100 arm. The current commercial processes use a fixed bed catalytic reactor in a gas recycle loop. There are a wide range of mechani- cal designs used to control the heat released from the reaction and the temperature profile so as to increase per pass conversion. Lurgi and ICI, Inc. (Imperial Chemical Industries) technology dominate. Other designs are offered by Mitsubishi, Linde, and Toyo corporations. New developments in methanol technology include the following: · Fluidized bed methanol synthesis being developed by Mitsubishi Gas Chemical. In this design a fine catalyst is fluidized by the syngas. Better contact between syngas and catalyst gives a higher methanol concentration

88 FUELS TO DRIVE OUR FUTURE exiting the reactor, which reduces the quantity of recycle gas, the recycle compressor size, and the heat exchanger area in the synthesis loop. · Liquid-phase slurry reactor for methanol synthesis. This DOE- supported effort is being developed at LaPorte, Texas, by Air Products and Chemicals, with technical assistance by Chem Systems, Inc. It is estimated that the Mitsubishi technology will reduce capital investment and natural gas usage by 7 and 6 percent, respectively. Methanol-Derived Fuels Methanol to Gasoline (MTG) Methanol can be converted to gasoline using the MTG process developed by the Mobil Research and Development Corporation. The Mobil MTG process is operating commercially in New Zealand, using the ZSM-5 cata- lyst in fixed bed reactors, to produce 14,500 bbl/day of gasoline. A flu- idized bed MTG reactor has been demonstrated by Mobil in a 100-bbl/day semiwork plant in conduction with West Germany. This program was jointly conducted by Mobil, URBK (Union Rheinische Braunkohlen Kraftstoff AG), and Uhde Gmbh, with additional funding from DOE and BMFI (the Bun- den Minister fuer Forschung und Technologie). Also, a direct heat-ex- changed MTG reactor concept has been developed by Lurgi. Haldor Topsoe and Mitsubishi have developed processes that are combi- nations of methanol (or other oxygenate) synthesis and MTG synthesis. The Advanced Mitsubishi Synthesis Gas to Gasoline process has been dem- onstrated at the pilot plant scale (1 bbl/day). The Haldor Topsoe process (called TIGAS-Topsoe Integrated Gasoline Synthesis), which is similar to Mitsubishi's process, is based on a mixture of methanol and dimethyl ether as an intermediate. The TIGAS process has been demonstrated at the semi- commercial scale in Houston. Methanol Conversion to Olefins and Diesel During the MTG development work at Mobil, it was discovered that the yield could be shifted to light olefins by varying the process conditions. The 100-bbl/day fluidized bed semiwork plant in West Germany was also used to demonstrate the methanol-to-olefins (MTO) mode of operation. High- quality gasoline is also produced. Lurgi has also reported MTO pilot plant results using commercial catalysts. Using olefins from the MTO process (and the F-T process described below) diesel and gasoline can be produced. Catalytic polymerization is a standard refinery process using acid catalysts and is currently being used at SASOL to convert C3-C4 olefins to gasoline and diesel. Recently, Mobil

CONVERSION TECHNOLOGIES A;IVD R&D OPPORTUNITIES 89 has developed an MOOD (Mobil olefin to gasoline and diesel) process using a Mobil commercial zeolite catalyst. The primary products are methyl- branched isoolefins, which have good octane ratings in the gasoline range. The diesel-range olefins can be hydrogenated to give isoparaffins, which have excellent diesel properties. A commercial-scale test of the MOOD process was successfully conducted at a Mobil refinery in late 1981 using feedstock from an FCC unit. Methanol for Electricity Generation Coal gasification is currently being developed to generate syngas for power generation. These facilities are referred to as integrated gasification combined-cycle (IGCC) power plants. Although syngas is used directly as a fuel, methanol can also be produced as a coproduct. The production of a methanol coproduct in an IGCC system could have the following advan- tages: (1) making available a clean liquid fuel for use in peaking service or for sale; (2) increase flexibility of service; and (3) level the IGCC plant at constant operation (95 percent on-stream factor) using methanol for cycling duty. F-T Synthesis and Product Upgrading The F-T process is a nonselective polymerization process that produces a range of products, including light hydrocarbon gases, paraffinic waxes, and oxygenates. Further processing of these products is necessary to upgrade the waxy diesel fraction, the low-octane-number gasoline fraction, and the large amount of oxygenates in the product water. Commercial F-T Processes Hydrocol Process (Hydrocarbon Research, Inc.~. The only U.S. com- mercial natural gas-to-liquids F-T facility was operated at Brownsville, Texas, from 1950 to 1957. The plant produced an olefinic gasoline with a small quantity of diesel fuel. The plant was shut down in 1957 because of the abundance of low-price crude. SASOL. The Lurgi/SASOL Arge fixed bed process has been operated commercially at Sasol I in South Africa for about 30 years and, using an iron catalyst, produces predominantly a waxy product. The Sasol Synthol circulating fluid bed process also uses an iron catalyst but operates at a higher temperature. Primarily light olefins and olefinic naphtha are pro- duced. This process has been operated commercially at Sasol I, II, and III. The naphtha requires reforming to meet fuel specifications.

9o FUELS TO DRIVE OUR FUTURE The upgrading scheme for Synthol products, as practiced at Sasol II and III, is complex. Major processing steps include a heavy catalytic polymeri- zation unit to upgrade C3-C4 olefins to gasoline and distillate, catalytic isomerization of the Cs-C6 fraction, hydrogenation and reforming of the C7- 190°C fraction, hydrogenation and cracking of the 190°C+ fraction to lower its pour point, and aqueous-phase distillation and hydrogenation of oxygen- ates to convert aldehydes into alcohols. The Sasol Synthol process has been selected for the Mossel Bay natural gas conversion project in South Africa. F-T Processes Under Development Shell. Shell Oil Company, in 1985, announced its SMDS (Shell middle distillate synthesis) process, which produces primarily middle distillate. Shell recently announced plans to use this process to convert natural gas to distil- late in Malaysia. The technology uses a fixed bed tubular reactor similar to the Arge design, with a more active catalyst believed to be a promoted cobalt catalyst. The syngas would be produced by a conventional partial oxidation process. The heavy, waxy product (mostly paraffins) is fed to a heavy-paraffin conversion reactor that uses a commercial Shell catalyst for hydroisomerization and hydrocracking to high-quality diesel. The naphtha product must be upgraded further. Slurry F-T Reactor Design. The slurry reactor, first used by Fischer in the 1930s, was demonstrated at the Rheinpreussen-Koppers demonstration plant in 1953. In the late 1980s a major development was conducted at the Mobil Research and Development Corporation with partial funding by DOE. The development involved the upgrading of a total vaporous F-T reactor effluent over a ZSM-5 catalyst. Other efforts in F-T technology include those by STATOIL, Gulf-Badger, Dow, Exxon, BP, Amoco, Mobil, and Union Carbide corporations. Economics The cost of syngas-based fuels, using reliable estimates, is between $45 and $62/barrel oil equivalent (10 percent discount rate; see Figure 3-1~. Syngas-based fuels from natural gas are expensive because of the high value of domestic natural gas for conventional markets. For example, natu- ral gas at $5/million Btu represents $33/barrel of the $60/barrel cost of MTG gasoline using the fluid bed reactor design. Although Alaskan natural gas would be significantly less costly, higher capital and transportation costs for liquid fuels produced in Alaska would offset the gas cost advantage. For example, estimates show that a natural

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 91 gas-to-methanol plant would cost 70 percent (California Fuel Methanol Study, 1989) more to construct at Prudhoe Bay than at a U.S. Gulf Coast location. Also, the cost of shipping methanol to Southern California from Prudhoe Bay is about $40/barrel oil equivalent ($22/actual barrel) compared to $7/ barrel ($4/actual barrel) from the U.S. Gulf Coast. These costs assume that methanol shipped from Prudhoe Bay would require a new pipeline to Val- dez and dedicated tankers from there to Southern California. Methanol from the Gulf Coast would require dedicated tankers. Because of methanol's lower energy density, greater volumes must also be shipped than would be the case for petroleum products. Even if syngas-based fuels from natural gas were to become viable ow- ing to a combination of cost reductions and special situations, exploitation would use foreign natural gas in a foreign location. At foreign locations, such as the Middle East, South America, or the Caribbean, natural gas would be significantly less costly than domestic gas because no local mar- ket exists and production costs are low. These foreign locations also meet the criteria of a reasonable construction cost environment and low transpor- tation costs to major world markets. For example, estimates show that a methanol plant costs only 10 to 25 percent more in the Middle East and South American or Caribbean locations than in the U.S. Gulf Coast. Trans- porting methanol from these sites to Southern California was estimated to cost only $4/barrel oil equivalent (California Fuel Methanol Study, 1989~. Syngas-based fuels from coal are expensive primarily because of high capital charges for the large capital investment required. For example, at a 15 percent discounted cash flow rate of return, the capital charges for mak- ing gasoline using the fluid bed MEG design represents $43/barrel oil equiva- lent of the $78/barrel total cost. The cost of syngas-based fuels from coal would be reduced if emerging technologies, such as the Shell gasifier and the slurry Fischer-Tropsch proc- ess, are used. Advanced gasifiers, such as the Shell gasifier, achieve a high thermal efficiency primarily by minimizing steam consumption. Syngas containing minimum amounts of hydrogen, i.e., having low hydrogen-to- carbon monoxide ratios, is produced. The slurry F-T process can use this syngas directly to make liquid fuels without further addition of steam or water; thus, maintaining high overall efficiency. Although advances in technology will reduce the cost of syngas-based fuels, it is unlikely that these cost reductions will reduce costs to $30/barrel primarily because the technologies are relatively mature. According to Schulman and Biasca (1989), the cost of methanol from coal would be reduced by $4/barrel oil equivalent due to improved synthesis gas unit de- signs. Using western coal in a fluid bed gasifier would reduce the cost of methanol by an additional $6/barrel. Methanol from underground coal gasification (UCG) costs less than from

92 FUELS; TO DRIVE OUR FUTURE surface gasification; that is, $44 to $57/barrel rather than $53 to $65/barrel. However, since the reliability of the estimate is considered speculative, conclusions about UCG cannot be made at this time. Conclusion and Recommendations for the DOE Program Processes to produce methanol or F-T liquids from syngas continue to be studied vigorously by industry, and methanol and F-T liquids may well find application in the United States. Production is, however, expected to be primarily outside the United States, where low-cost natural gas is available. While an interesting area of research in which further advances can be expected, the above factors discourage DOE research beyond that of funda- mental and exploratory research. DIRECT COAL LIQUEFACTION U.S. recoverable coal reserves are large, representing a significant pro- portion of world energy resources, and their prices are likely to remain modest (EIA, 1989a,b; Table 1-2; Table D-2. Unless a major new coal conversion industry is developed very rapidly, any shortage of coal is un- likely, the industry growth should not strain U.S. engineering or construc- tion capabilities, and the price advantage of coal as a feedstock over natural gas and petroleum will probably improve. Coal has additional advantages over other solid feedstocks that might be converted to liquid fuels. Because coal is geographically dispersed, its commercial use will produce less concentrated environmental impacts and more manageable demands on local infrastructure. Coal also has a high concentration of hydrocarbons, reducing mining, transportation, and proc- essing costs. There is also a well-established U.S. mining industry. Technology In direct liquefaction, hydrogen is added to coal in a solvent slurry at elevated temperatures and pressures. The process was invented by Freid- rich Bergius in 1913 and was commercialized in Germany and England in time to provide liquid fuels during World War II. The first U.S. testing of direct liquefaction processes followed World War II (Kastens et al., 1949~; efforts in the area declined when inexpensive petroleum from the Middle East became available in the early 1950s. Interest revived when the Arab oil embargo of 1973 caused high oil prices, resulting in increased federal funding for such research. A variety of process concepts were examined on a small scale, and three were tested on a large scale in the late 1970s and early 1980s: SRC-II (solvent refined coal) in Tacoma, Washington; EDS

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 93 (Exxon donor solvent) in Baytown, Texas; and H-coal in Catlettsburg, Ken- tucky. The DOE provided most of the funding for these successful demon- strations, but none proved economical as oil prices fell in the early 1980s. U.S. research continued after the big pilot plants were abandoned, most of it funded totally or in part by DOE. Few of the smaller pilot plants survived; today the only integrated pilot plant operating full time on direct coal liquefaction is the Advanced Coal Liquefaction R&D Facility in Wilson- ville, Alabama. Test units are available for contract at Hydrocarbon Re- search, Inc., Lummus-Crest, and the University of Kentucky, and Amoco Oil Company has smaller bench-scale pilot plants in operation. As indicated in this report, the projected cost for making liquid fuels from coal is about $40/barrel. This is about half of the approximately $60 to $80/barrel projected 10 years ago (Lumpkin, 1988~. Although many contributed to improving the technology, the Wilsonville pilot plant has become the focus for the U.S. direct coal liquefaction program, and esti- mates of the improved economics are based on the technology demonstrated there. The process at Wilsonville initially employed a single-stage dis- solver, followed by a filter to remove undissolved coal solids and a still to recover the solvent for reuse in the dissolver. The product was a solid with low ash and sulfur content to be used as boiler feed. The facility evolved in steps into a more versatile 6-ton/day liquefaction plant. The improvement in economics cannot be attributed to any single break- through but is rather the accumulation of improvements over several years of operation. First, a more effective and reliable process to remove solids from the liquid product by controlled precipitation replaced the filter. A second catalytic reactor was added to improve control over the chemistry of liquefaction. This reactor was first installed downstream of the solids re- moval and distillation systems; moving the reactor upstream further im- proved operation. Some of the recycled liquid used to slurry the feed coal was then bypassed around the solids removal unit, increasing the efficiency of the unit. Improved catalysts were added to both the first and second reactors. This series of modifications led to higher liquid yields, improved conversion of nondistillable liquids, less rejection of energy along with discarded coal minerals, and increased throughput relative to early two- stage systems. The success of this evolution shows that steady, long-range R&D can achieve major technological advances. The U.S. direct liquefaction process appears superior, but there is signifi- cant related activity overseas. The Japanese are operating a 50-ton/day liquefaction plant in Australia (Coal and Synfuels Technology, 1989b). Information on its operation is sparse, but apparently there have been some difficulties. The Japanese say they have developed a catalyst superior to any on the market, but samples are not available for testing. A 150-ton/day pilot plant, using a similar process, is being designed for installation in the

94 FUELS TO DRIVE OUR FUTURE 1990s in Japan. The West Germans have operated a 200-ton/day pilot plant at Bottrop since the early 1980s, but it was recently converted to the study of upgrading petroleum residuum. The West Germans have an unmatched backlog of experience in coal liquefaction on a large scale, although their process has some drawbacks compared to recent U.S. developments. The British are building a pilot plant about the size of Wilsonville in North Wales, at Point of Ayr. Their smaller-scale work was very encouraging, and this process may be a strong competitor if it works out on the larger scale. All of these projects are government funded. Fuel Properties Products of direct coal liquefaction are expected to meet all current speci- fications for transportation fuels derived from petroleum. Major products are likely to be gasoline, propane, and butane. Distillate fuels can be made but would likely require large volumes of hydrogen. Gasoline may be a particularly attractive product because it would have a relatively high oc- tane. High octane is achieved by the high aromatic content of the liquids. If regulations are established limiting the aromatic content of gasoline for environmental reasons, the cost of liquid fuels produced from coal by direct liquefaction would rise. While the benzene content of gasoline made from coal is extremely low, the concentration of other aromatics is high, and they could be hydrogenated to produce naphthenes at a moderate increase in cost. This would increase the volume of the products, decrease octane number, and increase hydrogen consumption. Environmental Considerations With regard to environmental emissions with local impact, a coal lique- faction facility is broadly comparable to a refinery with up-to-date emission control systems. As in many conversion processes, sulfur and nitrogen are removed from the feedstock and appear in the product fuels at greatly re- duced levels. Wilsonville has successfully shown that local emissions can be controlled satisfactorily (see Chapter 5 regarding greenhouse gas emis- sions). Direct coal liquefaction is fairly energy efficient; about two-thirds of the coal fed comes out as liquid product, and the rest is consumed to run the process. One area of concern is industrial hygiene. The intermediate products ot coal liquefaction (internal to the process plant) are polynuclear aromatic hydrocarbons, which are well-known carcinogens and mutagens. Industry and government programs over the past 20 years have demonstrated that proper attention to hygiene can make coal liquefaction plants safe places to work (U.S. DOE, 1989b).

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES Potential Cost Reduction 95 The improvements demonstrated at Wilsonville were incorporated in a new design that is an update of the DOE-funded Brechenridge Project, which used the H-coal process and was completed in 1981. The new design uses two reactors in series in place of the single-stage H-coal design and includes a simplified product distillation system, the new solids removal technology, and the simpler water treatment methods proved at Wilsonville; the process upgrades most of the liquid products to gasoline blending com- ponents (Lumpkin, 1988~. The result is a projected cost decrease of about 60 percent, to about $40/barrel (crude oil equivalent). Substantial improvements are still likely if research continues. Recent runs at Wilsonville have achieved a doubling of coal feed rates from that assumed in the above design by making the reactor temperatures more uni- form. These high rates may reduce costs by $2 to $3/barrel. Equipment is being installed to allow better distillation of the product liquids, which will in turn reduce the amount of high boiling components in the products. This change may reduce costs another $2/barrel. There is a large incentive to learn how to process different coals. Most recent Wilsonville trials used Illinois coal. Runs with a higher-rank, bitu- minous Ohio coal have demonstrated very high liquid yields, but the quality of the liquids needs to be improved. Lower-rank coals, either subbitumi- nous or [ignites, are relatively cheap and convert to high-quality liquids. However, their high moisture content and high levels of oxygen lead to problems that have not been entirely resolved at Wilsonville. Resolution of these problems could reduce costs by as much as $4/barrel. Removal of ash, and perhaps the unreactive parts of the coal itself, be- fore liquefaction could improve conversion, reduce erosion, and eliminate the need for the current solids separation process. Technologies are under development, funded primarily through DOE or the Electric Power Research Institute, that might be used to clean liquefaction feedstock, although they are primarily intended to prepare cleaner power plant fuel. Capital invest- ment at Wilsonville is needed to adapt these technologies and determine their economic attractiveness. There are clearly many opportunities to improve the economics of direct coal liquefaction. The DOE hopes to reduce costs at Wilsonville by 15 percent within the next 3 or 4 years. This target seems conservative. Direct liquefaction is capital intensive, and its total cost is relatively insensitive to most individual improvements. Multiple improvements are needed to significantly reduce costs. Such improvements are likely if re- search continues. Environmental requirements to reduce the aromatic content of gasoline may increase the costs of producing this fuel from coal. Although it is impossible to predict whether major technical breakthroughs

96 FUELS TO DRIVE OUR FUTURE will occur, the possibility should not be discounted. The high level of U.S. and foreign fundamental research on coal structure and chemistry for the past 10 years could lead to a superior means of conversion. DOE Program on Direct Liquefaction The recent assessment of research needs conducted by DOE's Office of Program Analysis outlines an excellent program aimed at bringing down the cost of direct liquefaction (Schindler, 1989~. Industry members of the as- sessment panel particularly stressed the need for federal funding of a large- scale pilot plant, as large as the German effort at Bottrop or the planned Japanese pilot plant (i.e., processing 150 tons per day or greater) to develop hardware and perform chemical studies. A broad range of fundamental and exploratory research was also recommended, based on the recognition that possible improvements in the current technology appear limited but that breakthroughs may bring down the cost of liquid fuels produced from coal to below $20/barrel. The committee concurs with these research recom- mendations. In between these two extreme types of development, intermediate-size flow units are needed. Wilsonville, or an alternative plant of about the same size, would be useful to test changes in process configuration at rea- sonable cost. Smaller pilot plants are also needed to test catalysts, explore operating conditions, and provide low-cost testing of new ideas. The com- mittee recommends that such small-scale work continue to be sponsored by DOE, with the work performed by private contractors in industry and uni- versities. In this way a wide variety of experts can contribute and technol- ogy transfer to industry will be enhanced. The general purpose pilot plant proposed for installation at the Pittsburgh Energy Technology Center is less attractive due to lower industrial participation. The DOE-funded programs that are relevant to the conversion of coal into transportation fuels in fiscal year 1990 allocate approximately twice as much money to process development as to each of the other categories (see Appendix G for definitions of fundamental, exploratory and catalyst, and process research). This emphasis on development may be unavoidable when industry is reluctant to participate because of the long time scale and uncer- tainty involved. The accuracy of the funding split is somewhat uncertain, since it is based on brief project descriptions that DOE provided the com- mittee. However, process demonstration, which is the step following proc- ess development, is receiving no funding, and without money for this pur- pose over the long run the United States will fall behind its foreign com- petitors. When the pilot scale demonstrates that processing 150 tons or more of coal per day has provided scale-up information, proceeding with a commer-

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 97 cial-scale demonstration facility should be seriously considered. This facil- ity should be a single-line plant with enough redundancy of critical equip- ment that an acceptable plant on-stream factor can be achieved. Only by operating a commercial-scale facility can its true economics be determined. It would be best if this commercial-scale demonstration could be fi- nanced on an international scale. Countries such as West Germany, France, Italy, and Japan depend on imported oil for transportation fuel. An interna- tional project funded by their governments and the United States on a cost- sharing basis, with participation of the private sector in the form of project management, engineering, and construction, operations, and maintenance, as well as private sector investment, would help ensure a successful effort. The Research, Development, and Demonstration organization of the Inter- national Energy Agency might coordinate and monitor the project. A suc- cessful commercial-scale demonstration would be valuable for the United States should it become desirable for the supply of liquid transportation fuels to be augmented through direct coal liquefaction technology. Conclusion Over the next 5 years research effort directed toward new catalysts and new processes should be stressed with a goal to selecting the best coal conversion processes for demonstration in a large pilot plant within this time frame. Achievement of this goal will require establishing technical confidence, achieving anticipated environmental requirements, and reduc- ing the cost so that industry is willing to participate. The program should include high-quality economic and technical evaluations by engineering firms, petroleum industry operating companies, and qualified consultants to guide the selections of the best technologies to move forward. COAL-OIL COPROCESSING Coal-oil coprocessing is a technology that simultaneously converts heavy petroleum residuum and coal to liquid transportation fuels. Incentives for coprocessing depend strongly on the existence of synergisms between the coal and resid as they are processed together. Coal may aid operability due to the solvency of coal liquids, and coal ash may scavenge metals from the resid to extend catalyst life. Other synergies may exist. In the 1970s the Canada Centre for Mineral and Energy Technology (CANMET) showed that the addition of less than 5 percent coal to a petroleum feedstock signifi- cantly improved distillate product yields (Rahimi et al., 1987~. This proc- ess was employed in a 5000-bbl/day plant started up in 1985 by Petro- Canada, near Montreal (Kelly and Fonda, 1984~. Background, state-of-the-

98 FUELS TO DRIVE OUR FUTURE art, and R&D opportunities for coprocessing technology are summarized in two recent reports (Schindler, 1989; Schulman et al., 1988~. The most significant disadvantage of using coal in petroleum upgrading is its impact on capital and operating costs—coal handling, hydrogen gen- eration, catalyst replacement, and waste amelioration. Recent Developments Various coprocessing technologies have undergone some development. Chevron Corporation ran a 6-ton/day coprocessing pilot plant at its Richmond, California, facility in 1983 (Shine et al., 1984~. Results reported for this close-coupled thermal catalytic system included good operability, a synergy for resid conversion, and demetallation of high-metal-content reside. In 1984 Kerr-McGee tested a process in which the bottoms from a resid hy- drotreater replace about one-half the recycle oil in a thermal catalytic two- stage coal liquefaction process (Rhodes, 1985~. Lummus coprocessing tech- nology includes two process flow schemes (Greene et al., 1986~. In one, coal and hydrotreated resid are fed to a two-stage process consisting of a short contact time (SCT) thermal reactor and an expanded bed LC-fining system. In the second scheme the resid is fed to the LC-finer only. In either case a solvent stream is recycled to the first stage, the product vac- uum bottoms are fluid coked, and the coke is gasified. Significant foreign developments have occurred in the West German 250- ton/day pilot plant operated by Veba Oil at Bottrop (Schulman et al., 1988~. It was used as a coal liquefaction pilot plant until 1986 and thereafter processed petroleum vacuum reside. Although not used for coprocessing, its operation with both coal and petroleum resid indicates the flexibility of the technology to accommodate different feedstocks. Current Developments The DOE is supporting the development of coprocessing in two pilot- scale programs at Hydrocarbon Research, Inc. (HRI), and UOP, Inc., and in various smaller-scale research projects. In addition, under the first round of DOE's clean coal technology program the department selected a 12,000- bbVday coprocessing project, using HRI technology, sited in Ohio. If this project is completed, it could mark the first large-scale demonstration of coprocessing technology. Further descriptions of the Signal-UOP and HRI technologies are given in Appendix H. In addition to these pilot-scale projects, advanced and fundamental coprocessing research is being conducted at U.S. universities and research institutions. Studies concern coal-oil interactions and process chemistry,

CONVERSION TECHNOLOGIES AlID R&D OPPORTUNITIES 99 with the goal of improving existing processes and identifying new process concepts. Some research on direct liquefaction and heavy resid upgrading may also apply to coprocessing. Fuel Properties Coprocessing is similar to direct coal liquefaction in that it produces fuels that are compatible with existing fuel markets. In particular, it is directed toward producing transportation fuels because these are the highest value-added products. Therefore, the emphasis has been on producing re- finery feedstocks and finished products that can meet motor and jet fuel specifications. The mid-distillate and vacuum gas oil products from coprocessing are low in sulfur and nitrogen (ruddy et al., 1986) and could be used as low-sulfur fuel oil or turbine fuel for utility applications. Environmental Considerations The aromatic nature of coal tends to impart aromatic content to the prod- ucts, which improves the octane value of the product naphtha for use as a gasoline-blending stock. Burning fuels from coprocessing will emit no more pollutants than their petroleum-derived counterparts. However, all fuels are produced at some loss in thermal efficiency and CO2 is produced. Other environmental im- pacts of coprocessing should be within the scope of existing petroleum refining and coal utilization practices. Opportunities for Cost Reduction Determining the existence and extent of synergism between coal and resid is needed to assess the economics. Several investigations place the product costs of coprocessing between those of heavy resid upgrading and direct coal liquefaction (Schindler, 1989; Schulman et al., 1988; Duddy et al., 1986; Huber et al., 1986~. Cost reduction will come from determining how to maximize the benefits of any synergisms that might exist. In par- ticular cases a combination of appropriate refinery equipment, resid costs, and coal availability might justify coprocessing. It appears unlikely that coprocessing will find application as a stand- alone technology. The economics of coprocessing require a significant difference between coal and resid costs to justify the additional capital expense to add coal to an existing refinery. Such a gap between coal and petroleum prices could justify construction of a coal liquefaction plant rather than a grass-roots coprocessing plant, to take greater advantage of coal as the less expensive feedstock. Of course, a coal liquefaction plant would

100 FUELS TO DRIVE OUR FUTURE probably be capable of coprocessing or resid hydrocracking if resid costs decrease. Other opportunities for cost reduction in coprocessing are the same as those in direct coal liquefaction: increased throughput, reduced hydrogen cost, lower coal costs, better catalysts, better engineering design, and iden- tification of new process concepts. Recommendation for the DOE Program Coprocessing of heavy oils or residuum with coal may offer an opportu- nity for the introduction of coal as a refinery feedstock. A demonstration plant for production of a clean boiler fuel is part of the DOE's clean coal technology program. Funding of basic bench-scale research should be con- tinued over the next 5 years to define the extent of synergism for coprocessing coal-resid combinations, followed by a thorough economic analysis quanti- fying the impact of this synergism. If favorable, the impact of synergism should be confirmed at the Wilsonville test facility to define optimum proc- essing conditions. If little or no synergism is found, work in this area should be terminated. COAL PYROLYSIS Description of the Technology Pyrolysis of coal dates back to the 18th century, using temperatures below 700°C in fixed or moving bed reactors. The primary product was a low-volatile smokeless domestic fuel, although the value of the liquid prod- ucts was also soon recognized. During the 1920s and 1930s there was a great deal of R&D in low-temperature processes, but interest died in the mid-1940s when gas and oil became readily available at low prices. With the oil embargo and increased oil prices of the early 1970s, interest renewed in coal pyrolysis, but in more recent times interest has again declined along with petroleum prices (Khan and Kurata, 1985~. In the most recent work, development was aimed at processes that maxi- mize the yields of liquid products. These processes require rapid heat-up, using fluidized or entrained bed reactors. A number of the processes re- quire the addition of reactants (steam, carbon dioxide, and hydrogen) at greater than atmospheric pressure to increase yields and limit secondary reactions. Reactor type, temperature, pressure, residence time, and coal type all have significant impacts on product yields (see Appendix I, Tables I-1 and I-2. Pyrolysis under mild temperatures (500° to 700°C) and pressures (up to 50 psi") with rapid heat-up can produce high liquid yields without adding hydrogen (hydrogen would have to be added to these liquids to produce

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES 101 transportation fuels). However, a significant part of the feed coal remains as char with less market value than the feed coal. As a result, coal pyroly- sis offers the promise of lower liquid costs only if the char can be upgraded to a higher-value product, such as form coke, smokeless fuel, activated carbon, or electrode carbon, or if the liquid yield can be significantly in- creased by using low-cost reactants (steam and carbon dioxide) or catalysts. Fuel Properties Different liquid fuel properties of the products of coal pyrolysis result from different processes. Processes include the Coalite process, a slow heating process that produces more gases and char than tar (Khan and Kurata, 1985~; the Occidental Research Corporation flash pyrolysis process, a rapid heat-up process that produces more tars (DeSlate, 1984~; and the FMC COED process (see Appendix I for properties). For the above order of processes the desirability of the char as a fuel decreases because of the increase in ash and sulfur content on a heat content basis and the deteriora- tion of its size consistency, which can lead to material handling problems. Pyrolysis liquids require extensive hydrogenation to be useful as trans- portation fuels. Another approach is to combine coal pyrolysis with pro- duction of synthesis gas to potentially increase the liquid yields for conver- sion processes producing transporation fuels. Environmental Considerations The environmental impacts of using liquid and char coproducts from coal pyrolysis will be very similar to those associated with the feed coal. Coal pyrolysis in the presence of alkaline material can result in the reten- tion of sulfur in the char with a corresponding reduction in the liquid prod- ucts. Care must then be taken to avoid uncontrolled releases of hydrogen sulfide from the char alkaline mixture (Gessner et al., 1988~. Economics The timing of a commercial application of mild pyrolysis will depend on the marketability of the char and the quality of the liquids. Spot market prices for metallurgical coke now exceed $130/ton. Assuming that form coke can command a comparable cost, mild pyrolysis could be economi- cally viable in the immediate future. Recommendations for DOE Research To realize the potential of mild pyrolysis as a source of transportation fuels, a number of research issues must be resolved. The DOE's current

102 FUELS TO DRIVE OUR FUTURE research program partly addresses these issues. Under its Surface Coal Gasification Program, research is focused on the development of advanced continuous mild gasification systems to produce optimal readily usable coproducts. The program is funding the development of four 100-lb/hour pilot units using fundamentally different approaches. The DOE should continue funding this program, with special attention to the quality of the liquids produced, the value and marketability of the char, and the size of the coal resource base that can be used with this technique. In the future DOE should attempt to scale up one or a combination of the most promising processes to obtain adequate design data for a commercial demonstration. The committee concurs with a number of recommendations on coal py- rolysis R&D made by an assessment panel to DOE (Schindler, 1989~. The first was to study the chemistry and mechanisms of catalytic hydropyroly- sis. Another important recommendation was to conduct a systems analysis of pyrolysis or hydropyrolysis coupled with gasification and combustion as a means of utilizing the char. DIRECT CONVERSION OF NATURAL GAS Recently, substantial research activities have been conducted in the area of natural gas (methane) conversion to methanol without the use of syngas (Kuo, 1984; Kuo et al., 1987~. The primary commercial goal of this re- search is to convert remote natural gas, which cannot easily be brought to market and is of little commercial value, into more easily transported liquid fuels. Except for Alaska, all of the remote low-cost gas is located in other countries. Technology and State of Development Numerous direct methane-to-methanol conversion routes are being stud- ied at the bench scale by various companies, government agencies, and universities. These include cold flame oxidation (direct partial oxidation) in which the main chemical reaction is the oxidation of methane to metha- nol, direct oxidation involving the catalytic coupling of methane and an oxidant to produce C-2 products and hydrocarbons, oxychlorination, indi- rect oxidation with oxidative coupling to ethylene, and catalytic pyrolysis involving contact of methane with a catalyst. ARCO Oil and Gas Company appears to be a leader in indirect oxidation, and recent success has been reported with its REDOX process. Other conversion routes include strong acid conversion and biological conversion (see Appendix J for additional technical details).

CONVERSION TECHNOLOGIES AND R&D OPPORTUNITIES Economics 103 Although the technology exists for converting natural gas into liquid fuels, the cost is too high when the technology involves first converting natural gas into syngas. The costs for syngas-based fuels from natural gas, expressed as 1988 dollars per crude oil equivalent barrel, depends on the technology and assumed rate of return and vary as follows: (1) methanol from natural gas, $45 to $50/barrel; (2) methanol to gasoline, $60 to 68/ barrel; and (3) the Shell Middle Distillate Synthesis process, $58 to $64/ barrel (see Chapter 3 and Tables D-3 and D-4. The direct conversion routes have the potential of being more energy efficient and less expensive since they bypass the formation of syngas. However, the current level of development has not achieved the potential significant cost reductions. Gasoline from ARCO's REDOX process costs more than Mobil's MTG (fluidized bed) process (Schumacher, 1989~. An analysis of the cold flame oxidation route, showed that based on an optimistic design the cost of gasoline would be reduced only 7 to 15 percent (with zero cost for natural gas) compared to the conventional MTG technol- ogy (Fluor Corporation, 1988~. This analysis also indicated that the cold flame oxidation route did not have any overall thermal efficiency advan- tage. Liquid fuels from domestic natural gas are expensive because of the high value of domestic natural gas for conventional markets. For example, natu- ral gas at $5/million Btu represents $33/barrel of the $60/barrel (10 percent discount rate) cost of MTG gasoline using the fluid bed reactor design. Even if a direct methane conversion process were developed that used 20 percent less natural gas, the cost of natural gas would represent $26/barrel of crude oil equivalent of the gasoline cost. Although Alaskan natural gas would be significantly less costly, higher capital and transportation costs for liquid fuels produced in Alaska would offset the gas cost advantage. Estimates show that a natural gas-to-metha- nol plant would cost 70 percent more to construct at Prudhoe Bay than at a U.S. Gulf Coast location. Also, shipping methanol to Southern California would cost about $40/barrel oil equivalent from Prudhoe Bay compared to $7/barrel from the U.S. Gulf Coast (California Fuel Methanol Study, 1989~. Even if liquid fuels from natural gas were to become viable owing to a combination of cost reductions and special situations, exploitation would use foreign natural gas. At foreign locations, such as in the Middle East, South America, and the Caribbean, natural gas would be significantly less costly than domestic U.S. gas because no local market exists and production costs are low. These foreign locations also meet the criteria of a reasonable construction cost environment and low transportation costs to major world markets. For

104 FUEI~i TO DRIVE OUR FUTURE example, methanol plant costs are only 10 to 25 percent more in Middle Eastern, South American, and Caribbean locations than on the U.S. Gulf Coast. Transporting methanol from these sites to Southern California was estimated to cost only $4/barrel oil equivalent (California Fuel Methanol Study, 1989~. Recommendations for the DOE Numerous direct methane conversion routes are being studied at the bench scale by various companies, government agencies, and universities that avoid the need to produce syngas as an intermediate. These direct conversion routes have the potential of being more energy efficient and less expensive since they bypass the energy-intensive and expensive step the formation of syngas. However, the current level of development has not achieved the potential significant cost reductions. Even if liquid fuels from natural gas were to become viable owing to a combination of cost reductions and special situations, exploitation would use less valuable foreign natural gas in a remote location. Therefore, gov- ernment-sponsored research on direct methane conversion technology should be limited to fundamental research.